aiqbench analysis
This study reveals a critical generalization gap in human activity recognition models when shifting from younger to older adult populations, demonstrating that standard benchmark improvements often fail to transfer effectively across age groups. The research indicates that incorporating diverse pretraining data, such as the UK Biobank dataset, significantly narrows this disparity by providing richer representations that better capture population diversity. Consequently, advancing wearable health monitoring requires moving beyond architectural scaling alone toward strategies that explicitly account for demographic differences in movement patterns.
Characterizing the Performance Gap in Human Activity Recognition for Older Adults
arXiv:2610.02711v1 Announce Type: cross Abstract: Human activity recognition (HAR) from wrist-worn accelerometers is increasingly used for health and behavioral tracking. Yet, most wearable HAR models are developed and evaluated on datasets dominated by younger adults, leaving it unclear whether benchmark progress generalizes across age groups. In this work, we leverage MyMove, our carefully annotated, free-living older-adult HAR dataset (mean age 71), to evaluate deep-learning architectures and training regimes under both leave-one-subject-out and cross-dataset transfer. We find that improvements on younger-adult benchmarks fail to transfer equally to data collected from older adults, resulting in a persistent and often widening performance gap. However, richer representations, particularly frozen self-supervised features pretrained on the age-diverse UK Biobank dataset, substantially improve performance on data from older adults and consistently narrow the performance gap, at modest cost to younger-adult performance, though disparities remain. These findings suggest that benchmark gains and architectural scaling alone provide an incomplete picture of progress in wearable HAR, and broader advances may require representations that better capture population diversity, alongside personalized adaptation to individual movement patterns and routines.
Abstract:Human activity recognition (HAR) from wrist-worn accelerometers is increasingly used for health and behavioral tracking. Yet, most wearable HAR models are developed and evaluated on datasets dominated by younger adults, leaving it unclear whether benchmark progress generalizes across age groups. In this work, we leverage MyMove, our carefully annotated, free-living older-adult HAR dataset (mean age 71), to evaluate deep-learning architectures and training regimes under both leave-one-subject-out and cross-dataset transfer. We find that improvements on younger-adult benchmarks fail to transfer equally to data collected from older adults, resulting in a persistent and often widening performance gap. However, richer representations, particularly frozen self-supervised features pretrained on the age-diverse UK Biobank dataset, substantially improve performance on data from older adults and consistently narrow the performance gap, at modest cost to younger-adult performance, though disparities remain. These findings suggest that benchmark gains and architectural scaling alone provide an incomplete picture of progress in wearable HAR, and broader advances may require representations that better capture population diversity, alongside personalized adaptation to individual movement patterns and routines.
Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
arXiv:2610.02711 [cs.HC]
(or arXiv:2610.02711v1 [cs.HC] for this version)
https://doi.org/10.48550/arXiv.2610.02711 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
https://doi.org/10.1145/3830727.3834835 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 278 words
Related coverage
AI Companion Robots Are Closing the Human Connection in Modern Homes
IEEE Spectrum · Aug 25, 2026
Unsupervised Keypoints for Real-Time Fall Detection: Comparative Analysis Under Real-world Conditions with Predictive Bandwidth Reduction
arXiv · Jul 21, 2026
Unsupervised Keypoints for Real-Time Fall Detection: Comparative Analysis Under Real-world Conditions with Predictive Bandwidth Reduction
arXiv · Jul 20, 2026
Digital States 2026: AI Moves From Experiment to Infrastructure
govtech.com · Sep 24, 2026